CC Meeting Full 20070320
hengsin [n=chatzill@60.51.88.186] ha entrado en la sala.
trifon [n=chatzill@home-213-240-227-143.megalan.bg] ha entrado en la sala.
crooney [n=crooney@89.19.69.9] ha entrado en la sala.
crooney: hi guys what's the topic today?
CarlosRuiz: Hello
AFalcone2: Hi all
teo_sarca: hello all
trifon: Hi
teo_sarca: what are the topics for today ?
hengsin: hi all
CarlosRuiz: firstly review the 3.1.5 status
teo_sarca: hi hengsin
hengsin: there are still some open priority 9 bugs and no in deep functional testing yet
hengsin: which mean we 've slip from the previously set target date
teo_sarca: hmmm, yep there are 12x9 priority
trifon: who sets priority to 9 of this? 1680496 - Performance/Dashboard tab is basiclly unreadable
CarlosRuiz: I did, is commented
trifon: ok.
teo_sarca: 8 x 9 priority are not assigned...
crooney: I'v ebeen worknig on the project ones ... but the whole area lots like it's riddled with problems ... although I haven't used it much before ... I just assume the comiere partners would have been using it and it was stable ... feels more like beta
crooney: lots=looks
CarlosRuiz: I thought the C_ProjectType_ID error was solved
CarlosRuiz: very before
crooney: well it's still there now!?
crooney: perhaps it was solved somewhere else?
CarlosRuiz: it's in Project window?
crooney: i'm using trunk to test!
crooney: yes
hengsin: I think it is working now although there is a message printed on console
CarlosRuiz: http://sourceforge.net/tracker/index.php?func=detail&aid=1618423&group_id=176962&atid=879332
crooney: ah yes it does set a value
crooney: just didn't like the error
crooney: popping up
hengsin: croo: are u sure it doesn't work ? AFAIK, it is just a warning
CarlosRuiz: oracle or postgres?
CarlosRuiz: AFAIR even the warning was solved
crooney: oracle, like I said it sets a value ... but still getting a beeb and a message on the console
hengsin: the beep is from eclipse
crooney: i will check the link
crooney: eclipse? :) I will check again
crooney: i set the priority low anyway ....
crooney: i was looking more at the gl postings reported in 1564466
crooney: and 1564456
hengsin: [1634461] should have been resolved, just waiting victor/moyses to confirm..
hengsin: [ 1632875 ] Re-activate SO does not unreserve stocks - anyone tested this ?
crooney: i set the priority but I didn't test the supplied solution
CarlosRuiz: there are some unprioritized
CarlosRuiz: priority = 5
crooney: i think they are new? I left one in spainish ...
crooney: hmm no some are old
crooney: below 25 ... I guess that's Mike
hengsin: ok, what shall be the next step now ?
CarlosRuiz: ok, just commented the "set project type"
crooney: ok thx .. and yes the beeb was eclipse :)
crooney: or maybe not I have too many windows open
crooney: anyway I guess we just continue targeting the pr 9 bugs!?
hengsin: can we set a new target date for 3.2 ?
CarlosRuiz: there are 12 priority 9 - and 24 priority 5 - unprioritized
crooney: and 1 8?
crooney: ok I can loko through the remaining 5s ... I'm not going to test just evaluate and prioritised based on the report
CarlosRuiz: but phib bugs are from Compiere 253b, mostly solved
crooney: hopefully!
hengsin: we need to test those ..
crooney: phib was the project bugs I was looking at .. I didn't see anythnig changed!
crooney: if they are fixed that wil be great because they are mostly 9s
crooney: ah you mean the phib 5s
CarlosRuiz: [ 1675206 ] Error when I try Merge Entity
CarlosRuiz: is incompletely solved but closed
hengsin: first is to cross check whether it is fixed in Compiere
CarlosRuiz: must I reopen [ 1675206 ] Error when I try Merge Entity ??
CarlosRuiz: is partly solved - or must I open a new bug?
crooney: well if it's solved why re-open?
CarlosRuiz: not solved completely
CarlosRuiz: I mentioned three programs with the same problem, just one fixed
crooney: if its the same problem re-open .. if it's different open new ??
CarlosRuiz: ok
crooney: sounds like a reopen then ... I guess what ever is easiest ... so long as we know what was fixed and how I guess
hengsin: victor close it. however the other 2 sounds like separate bug report, i.e not related to 'merge entity'
CarlosRuiz: ok, reopened, now we have 13 priority 9
teo_sarca: what you think about http://sourceforge.net/tracker/index.php?func=detail&aid=1619300&group_id=176962&atid=879332
teo_sarca: ?
crooney: of course when detailed testing happens we may end up with more 9s
crooney: yes I looked at this briefly ... i think it's important but someone did offer a good work around ... in practise probably only Adim will have access to this window anyway?
crooney: would be nice if context dislpay too account of the window & tab details to decide what should be displayed or not
crooney: too=took
trifon: i have a patch for password fields
crooney: as carlos says in the post ... it should use the encrypted flag as defined in the AD
trifon: and credit card numbers
trifon: but it will take me some time to commit it.
crooney: any field with encrypted flag set trifon?
teo_sarca: trifon, the patch is using ad_column/field properties ?
trifon: not any just password. problem is that when encrypted chars increase so we need to increase length
trifon: of fields.
trifon: in fact any field which is set as pasword.
trifon: but admin must take care to increase size of fields.
trifon: also for credit card number i had to add new field which stores last 4 chars.
CarlosRuiz: ah, you're talking about encryption, not the context variable shown
trifon: yes, but the main problem i think is encryption.
trifon: if encryption is done then context will be encrypted too.
crooney: well that sounds like another problem!
trifon: looks so.
teo_sarca: but in context the fields are not encrypted... as i know
crooney: well the passwords are encrypted but displayed in the context!
CarlosRuiz: yes - "Encryption not working" -
CarlosRuiz: yes, I think in context will be shown as well, AFAIR they are decrypted when reading from DB
trifon: in fact there is no reall encryption in Adempiere at the moment.
crooney: yes
crooney: yes they are simply not displayed on screen ... but are stored open in the DB
trifon: i'm not sure how companies use Compiere/Adempiere in Germany as such behaviour is prohibited
trifon: but this is their problem.
CarlosRuiz: which behavior? unencrypted passwords in DB?
trifon: also enryption which allow decryption is not reall encryption.
trifon: yes.
CarlosRuiz: maybe because Oracle can solve the problem
trifon: how?
crooney: well I think it's extremely common that
crooney: not that it's good mind....
CarlosRuiz: in 10g there are some feature to encrypt columns, or even complete db
CarlosRuiz: virtual private database
crooney: i think for now if we can use the column/field settings used in te display to be the same in the window and the contect display ... that enough for this bug anyway
CarlosRuiz: yes, and open a bug for "Encryption not working"
CarlosRuiz: I remember encryption worked in some past version
crooney: yes ...
trifon: why do you think so Carlos?
trifon: even if column is set as encrypted Compiere will throw error as size of chracters do not fit into the column lenght.
CarlosRuiz: I remember that in some version (maybe 253a) I put ad_user.password as encrypted and when saving Compiere write in db encrypted password
CarlosRuiz: the problem I confront that time is that I couldn't access server monitor with encrypted pwd in db
trifon: yes login must be changed also.
trifon: as at the moment it check for plain text match, but it must check match of hashes.
CarlosRuiz: but swing login were working, was server monitor in web not working
trifon: this patch modify 2-3 classes bot sure for the exact number.
CarlosRuiz: hengsin, about 1634461
hengsin: yes, I've modify VInvoiceGen and VInoutGen to use t_selection which should have resolved 1634461
CarlosRuiz: yes, just wondering if Pay Selection has the same problem, but looks like it doesn't
CarlosRuiz: must we set another target date, or just wait until priority 9 is solved?
server_ [n=server@189.142.83.124] ha entrado en la sala.
server_: Hi
teo_sarca: hi
server_ ha salido de la sala (quit: Remote closed the connection).
crooney: i expect as the push for testing increasing we will find more pr 9s - so it might be pointless to set a date? I wold say when there are no more 9s AND it has been tested then we declare stable
crooney: but we must be careful ok not to be forever fixing 9s :)
CarlosRuiz: yes, I think a 3.1.6 - release candidate can be released
hengsin: agreed.
CarlosRuiz: and 3.1.7, 3.1.8 ... if new bugs-p9 appears ... and so on
crooney: ok - we want people testing with as many fixes implemented as possible so yes that makes sense
CarlosRuiz: or it can be a moment where we just release patches.jar, maybe instead of 3.1.7 (if not changing db or changing libraries)
CarlosRuiz: do you know if Johannes implemented the jasperreport add-ons ?
CarlosRuiz: I think it will be good to allow that before releasing 3.1.6
crooney: i think so .... what of the RMA & dropship enhancements ... some asked if that will be in 3.2 .. on the one hand it is new functionality .. on the other it will be new to completed half finsihed work already there?
trifon: do you know if Johannes implemented the jasperreport add-ons ?
trifon: ??? ???.
trifon: not yet
trifon: JasperReports will not work when started from java web start.
CarlosRuiz: ok, are we going to allow him to integrate that before 3.1.6?
crooney: no? i thought so but not with ALL the bells and whistles
crooney: hmm ok that seems like a problem
DeathMeat [n=DeathMea@adempiere.reisinger.de] ha entrado en la sala.
DeathMeat: hi all
trifon: hi
crooney: hi dm
teo_sarca: hi
hengsin: hi dm
DeathMeat: ok, let me start explaining what i implemented and why it is useful to have it in trunk
DeathMeat: the first thing are so called silent processes
DeathMeat: in the show help dropdown of a process you have now the option Silent process - take defaults
DeathMeat: if you select this one, and the process has parameters it will take the default logic for these parameters and whenever you launch the process (a jasper process or a normal one from menu or a button) it will immediately start the process
DeathMeat: this allows getting rid of annoying dialogs where you jsut a a checkmark before and after the process
DeathMeat: especially useful for jasper but can also be applied to any other process
DeathMeat: that's it for silent processes, what do you think?
kontro [i=tko@e82-103-192-111.elisa-laajakaista.fi] ha entrado en la sala.
DeathMeat: is it worth to have it in trunk?
CarlosRuiz: I think yes
trifon: it sounds good as a feature.
vpj-cd [n=Horus@189.142.83.124] ha entrado en la sala.
vpj-cd: hi
DeathMeat: and it does only apply if you set the process to be a silent one
trifon: hi
DeathMeat: hi victor
teo_sarca: hi victor
vpj-cd: I sorry I have issue with my connection
CarlosRuiz: is the webstart+jasper solved too?
DeathMeat: it is
DeathMeat: and already commited, since it was a bug
DeathMeat: everything well documented as tracker item,
trifon: + subreports?
vpj-cd: what is the subject
DeathMeat: i already created a tracker item for silent processes, if it is ok then i commit my work and update the tracker item
DeathMeat: yes trifon also subreports work like a charm
DeathMeat: tested very heavily with java 5 and 6, locally and via webstart
trifon: ok. good i expected to receive some update from the tracker, strange.
DeathMeat: and local reports as well as webdeployed ones
trifon: where are reports uploaded?
trifon: on web server?
CarlosRuiz: [ Colin - as you pointed - projects have lots of issues : https://sourceforge.net/tracker/?func=detail&atid=410216&aid=1684303&group_id=29057 ]
DeathMeat: whereever you want, local filesystem, fs share, samba, webserver (apache), appserver
trifon: ok. thank'si will check it also.
DeathMeat: ok
DeathMeat: i'm going to like the way of deploying using webApp
DeathMeat: ;)
CarlosRuiz: what is the other pending point of Jasper?
DeathMeat: yes the other one is:
CarlosRuiz: did you commit the document format?
trifon: Johannes what do you mean webApp?
DeathMeat: i mean, using webApp is a nice way of deploying reports
trifon: i agree.
DeathMeat: sometimes better than having it on a servers filesystem
crooney: webStart or webApp?
DeathMeat: ok, the other thing is, that i integrated jasper reports also for normal reports
DeathMeat: webApp
vpj-cd: what happen if we set the report into db
vpj-cd: as attachmnet
DeathMeat: now it is possible to have reports treated as jasper reports when the printformat has a jasper process behind
vpj-cd: we can set the files of jasper into db
DeathMeat: ehm, won't work out of the box vic
CarlosRuiz: I suppose it will work ok attached in database, but problem with subreports
DeathMeat: i guess the reports in database thingy is a separate topic and a RFE
DeathMeat: +worth
crooney: still so long as we know the limits ... we are still much better off than before! and certainly much better off the compieer!
DeathMeat: but back to my second modification, the ability of having JR as normal reports
DeathMeat: it is awesome! because, you all know if you have multiple reports for the same window/table a small popup appearsa displaying the different reports
DeathMeat: and this alsow orks with jasper
DeathMeat: the classes i extended a little bit are APanel and AReport, nothing serious only a few lines of code
trifon: is this commited?
DeathMeat: so Reports are no aware of a (perhaps) existing jasper process behind the printformat
vpj-cd: but you can upload any files into attachment
DeathMeat: no, like i said i need the OK since we have a feature freeze
DeathMeat: but i created a tracker item with source attached
CarlosRuiz: I vote for including that, and then release 3.1.6
DeathMeat: carlos review it (or at least he told me)
trifon: me too.
teo_sarca: me too
trifon: my vote is to include it.
hengsin: agreed.
CarlosRuiz: ok Johannes, you have to "go" :-)
DeathMeat: sounds great!
DeathMeat: thanks
CarlosRuiz: thanks to you
CarlosRuiz: Colin, my opinion is that RMA & dropships must be integrated after 3.2 - but opinion can change if they're solved before ;-)
AFalcone2: i'm back
AFalcone2: Johannes, we have pending the custom datasource too
AFalcone2: pointed by hengsin
AFalcone2: I will be free in 2 days to work with you
AFalcone2 ahora se llama AFalcone
DeathMeat: oh great
DeathMeat: yeah if you have time we will do a separate meeting and talk about how to increase security by using jasper
DeathMeat: the custom datasource is a step in the right direction
DeathMeat: definitely
AFalcone: yes Johannes, it's a good point
DeathMeat: if jasper/the datasource is aware of roles and permissions it would be really great
DeathMeat: also scriptlets are a important point in jasper
DeathMeat: scriptlets give you the ability to execute external java code
DeathMeat: for data gathering and also transformation
DeathMeat: so we could spawn a lightweight adempiere client in the background responsible for feeding the report with data
DeathMeat: whatever, just my initial thoughs, and curious on your ideas Alejandro
DeathMeat: just gimme a shot whenever you have time to discuss this in detail, preferred in the afternoon
CarlosRui1: ok, there are other points that Hengsin proposed to discuss, but it's late, I think those must be theme for another meeting, but we can start the thinking and chatting
CarlosRui1: maybe open a wiki page to start the thinking
CarlosRui1: the points are:
CarlosRui1: - roadmap
CarlosRui1: - TANG project as R&D lab
kontro: Sounds sane.
hengsin: I think we need more opinions for a proper road map
vpj-cd: what about of AJAX
hengsin: and to set a clear direction how the product/platform would evolve ...
CarlosRui1: currently there are many proposals - rule engine - ajax client - 3 tier architecture - JSF
kontro: Tang becomes easier issue talk about when there is some code... Working on it. I think that we can have interesting discussion about tang at Berlin.
hengsin: at the same time hightight on going experimental effort so people interested at those area can participate.
hengsin: hightight=highlight
CarlosRuiz ha salido de la sala (quit: Read error: 145 (Connection timed out)).
CarlosRui1: yes, and the initial roadmap must be revised
CarlosRui1: and we must review if TANG is the next generation or a lab
Vd. ahora se llama CarlosRuiz
kontro: Took so much longer to compare apis and products than I had thought.
kontro: I am not very good at taking the easiest path :)
CarlosRuiz: I read Kontro is thinking now on TANG as the R&D lab of Adempiere, and it sounds good, only big companies have such lab
crooney: perhaps we need another name rather than roadmap.. it strikes me that a bazaar driven project like this does not have any difinijtive route .. just a final destination
CarlosRuiz: agree, good idea Colin
CarlosRuiz: roadmap is a name that apply most for proprietary products
AFalcone: agee Colin
CarlosRuiz: we can have directions, policies
hengsin: yes but roadmap is a commonly accepted/understand term, think new comer ...
vpj-cd: yes I the issue with a new rooadmap is
crooney: well it something to think about ... perhaps on sematics but people see roadmaop and think something made in stone....
kontro: Soon (maybe tomorrow) I can really start testing xforms in practice. It is very interesting since it provides ajax and fatclient in same package.
vpj-cd: we need a form to migrate easy
crooney: where as we just have aspertaions
CarlosRuiz: IMHO roadmap is confusing people
vpj-cd: because Carlos?
vpj-cd: why Carlos?
CarlosRuiz: i.e. 3.3.0 -> fyracle support
crooney: or there is a lot of concentration on it ... people think "I'll wait until such and such is included"
CarlosRuiz: exactly
crooney: or keep asking dates for this or that
CarlosRuiz: people read the roadmap, and then ask us, when will we have 3.6.0 ?
kontro: Only direction should be set.
crooney: yes
trifon: i agree.
crooney: yes
CarlosRuiz: i.e. 3.6.0 have human resource - but I think Victor is planning to release earlier
crooney: so we want Db independence ... we want a web UI ... etc
vpj-cd: yes but how set fyracle in rooad map
CarlosRuiz: we must drop fyracle until a maintaner appears
kontro: Like: Adempiere is trying to became more like standard J2EE5 application which provides good interfaces for integration.
vpj-cd: ok we only need update the roadmap
crooney: kontro, or even kontro we address the business needs that going to JEE5 will address
CarlosRuiz: yes, roadmap must not match with any specific technology or brand or name, only in very special cases
CarlosRuiz: I mean, we must not name GWT in roadmap, cos you saw, something did very fast with ZK
vpj-cd: ok but ie 3.6 we are coplaice this roadmap
kontro: We have to promise that in future adempiere is easier to learn and manage with knowledge java developer/administrator has.
vpj-cd: yes but zk no use same programing form that swing
vpj-cd: you use xul another language
crooney: i think it's mostly just a naming issue for us .. but others see roadmap and think something very specific
vpj-cd: also I think gwt have more future
vpj-cd: because google as company is back
CarlosRuiz: but we must not close doors
hengsin: there is not problem of having a roadmap, we are just talking about the problem with the 'current roadmap' we have
CarlosRuiz: if someone can do that in GWT, wonderful, but if someone can do that in ZK, wonderful too
vpj-cd: in the future very tools to desktop will be supply to google
CarlosRuiz: if we state in roadmap -> GWT, then people contributing with ZK can feel like closed doors
vpj-cd: the future excel and word , are of google
crooney: i think the point is we should address the business requriements ... we state those and then we can chose the technology when the time arrives
kontro: Not google open standards.
hengsin: and of course, we should not put version number, technology etc into the roadmap for things that is not firm yet, probably just some direction for possible future enhancement for those stuff.
vpj-cd: yes I think that web is very importnat
vpj-cd: to our merket
vpj-cd: now very ppl are with openbravo because have a solution to web
kontro: People are in trouble with microsoft and nobody wants to give similar power to google.
CarlosRuiz: agree Hengsin, the roadmap name is not the problem, the problem is having target versions, and specific technologies
vpj-cd: adempiere have not
kontro: I think that GWT is not best solution for applications which need GUI created dynamically.
kontro: It has same difficulties than Swing.
vpj-cd: because
crooney: yes look at what we have now and how does that compare to what the roadmap says for 3.2?
CarlosRuiz: Ashley was very fast, and L&F is really good!!!
vpj-cd: but we have solve all issue current use swing
vpj-cd: in adempiere
kontro: Hmmm... Am I beating dead horse again :)
hengsin: guys, no point debating technology choice here, you need to experiment and benchmark, talk is useless :)
vpj-cd: yes but we can use any technology
CarlosRuiz: exactly, but the point is we must not state GWT in roadmap
CarlosRuiz: as we must not state fyracle
kontro: Yes there is surprising problems around every corner in every tech.
vpj-cd: but I think we should use that have more future
CarlosRuiz: I think we must allow bazaar show the power
vpj-cd: we have not use python only because is good
hengsin: fyracle is really an exception here, there was a maintainer which have left the project.
vpj-cd: we use java because is exelent and have good company as back
CarlosRuiz: and because JJ used it :-)
crooney: look at this stage we are doing the research and testing possible solutions ... because we know a web UI is implortant to many customers. So we say we will work to develop that but how it will finally materialise we do not need to say and business people do not care anyway
kontro: I think that ajax toolkits are only patches for couple years period until xhtml2, xforms and svg are supported in browsers.
vpj-cd: yes aldo the bank
vpj-cd: oracle
vpj-cd: sap
vpj-cd: all use java
vpj-cd: only some lost use .net :-)
DeathMeat: hehe
DeathMeat: agree
vpj-cd: but ok we can set out mind in roadmap
crooney: .net is very very very popular here in ireland :( the second home of MS
CarlosRuiz: :-D good point - lost
CarlosRuiz: Bill visited Colombia this week :-)
crooney: bye the way kontro Swings works with .net too :D - lol only joking with you!
hengsin: kontro: you can't always predict the future ... MS technology is not the best but they have dominated the desktop ...
CarlosRuiz: he offered "investment" here - that means -> extend monopoly
AFalcone: hehe Carlos, you have the George W. and Bill visit there!
vpj-cd: what do you think road map to 3.1.0
vpj-cd: also in Mexico
crooney: did you finger print him at the airport when he arrived? :)
vpj-cd: GW are review you back place :-)
AFalcone: yes Victor, they not visit us
CarlosRuiz: :-D
AFalcone: :)
CarlosRuiz: ok, this raised another question for me
CarlosRuiz: good for CC
vpj-cd: yes GW was in Merida
vpj-cd: also :-)
hengsin: I think what is important is to send a clear message/picture to the community, what is the committed item, what are being investigate and what is 'future' ( i.e don;t know when )
CarlosRuiz: what do you think on ZK client POC - vs - JSF proposal from Vince?
kontro: As IT industry matures road goes towards open standards. When standard implementations are good enough. There is no need to take any risks.
CarlosRuiz: I want to know opinion about 3-tier JSF
vpj-cd: I think zk use xul is xml language, but this are into server
kontro: That why I think betting on standards is safer than choosing most innovative technology.
hengsin: JSF is heavy on server memory usage
crooney: i think it's a big step!
vpj-cd: the engine to zk is into server this is problem , it is similat o echo2
kontro: Plus side is that there is more coders who know JSF than Zk.
hengsin: victor, no, zk be program using xul or using java api like swing
CarlosRuiz: my worry is that 3-tier JSF is difficult to implement - it needs core changes - what do you think?
crooney: yes... as I said i think it's a big step!
vpj-cd: yes Low but you use the server to generate the html and java script
hengsin: not really, JSF can also be implemented on top of currrent core, in fact, compiere is doing just that.
crooney: risky to do all in one go
vpj-cd: I think the tecgnology as seems have more future as jsp
CarlosRuiz: and possibly JSF will delay the web client, that can be quickly released with ZK
crooney: ok but that is not what vince proposed is it?
vpj-cd: is more easy
trifon: i think that we must have this disucssion in frums.
trifon: ZK vs. JSF.
trifon: more eyes more opinions.
vpj-cd: and GWT :-)
CarlosRuiz: I liked Rod's words shared by Colin: "don't change anything for technology's sake"
crooney: why can't we do both?
crooney: and gwt
crooney: if we get the core right we can imlpement any number of UI
CarlosRuiz: the point is not ZK vs JSF
DeathMeat: the only reason why i don't like/used GWT in the past is the lack of widgets, what about sortable tables?
vpj-cd: I think the main issue to use any tegnologya is development need leanr another language
CarlosRuiz: the point is proposal of a real 3-tier architecture for JSF client
vpj-cd: this is small productive
DeathMeat: i mean, who want's to write a custom renderer for such basic functionality?
vpj-cd: if we use echo2 or another you do not need learn new language or script
vpj-cd: you use3 java
CarlosRuiz: call to order :-) as Hengsin said, it's not time to discuss which technology is better, this only distract us
trifon: my opinon is that any eb part must be implemented after 3 tier architecture is done.
vpj-cd: ok
CarlosRuiz: what we're trying to discuss is policies
vpj-cd: the we talk about the roadmap
vpj-cd: Goals
vpj-cd: * Find and fix bugs
vpj-cd: * User interface improvement
vpj-cd: * Run on 100% opensource stack
vpj-cd: Release would include the following components:
vpj-cd: * Compiere 260a
vpj-cd: * Plugin and Migration Management.
vpj-cd: o Proposal:
vpj-cd: + Create a Tool to do migrations based on XML Formats, so that version migration will be easier.
vpj-cd: + Improve the tools created by Marco, Trifon and Rob (ADCK, 2Pack)
vpj-cd: + Use the Apache(http://db.apache.org/ddlutils/) DDLUtils libraries for this purpose.
vpj-cd: o Requirements
vpj-cd: + Application Dictionary maintenance
vpj-cd: + New version migration
vpj-cd: + New components integration
vpj-cd: + Patch Management
vpj-cd: + Export and Import the Data Dictionary
vpj-cd: + Export and Import Entity Data (Client)
vpj-cd: + Delete entities (Client)
vpj-cd: + Wizard to create new windows
vpj-cd: * Free PDF
vpj-cd: * Jasper Report integration
vpj-cd: * Oracle & Oracle XE
vpj-cd: * PostgreSQL support
vpj-cd: * Migration Script 253a to 260a
vpj-cd: Adempiere Tasks
hengsin: victor, stop please ...
vpj-cd: I think the goal to 3.10 is complait
hengsin: this is not wiki page!
CarlosRuiz: we're not talking about 3.1.0, neither 3.2.0
AFalcone: hehe
CarlosRuiz: we're talking about the future releases
vpj-cd: ok the what section the road map
vpj-cd: we need talk
CarlosRuiz: about the message to the community if we set up target goals for 3.4.0, 3.6.0 etc
CarlosRuiz: the message to the community if we state GWT as the tool
CarlosRuiz: this is a bazaar
hengsin: I think it should be something like this - Ajax Client ( Under Investigation ) - ZK ( Posterita ), GWT ( Victor ), JSF ( Vince ) etc
CarlosRuiz: people will bring us the best
CarlosRuiz: web client
vpj-cd: ok I am agree
crooney: yes ... just say web client
vpj-cd: but I think the web client is issue very important
hengsin: a clear message - what is committed ( i.e confirm ), what is under investigation ( experiment welcome ) and what is don't know when ( volunteer welcome )
CarlosRuiz: very important indeed
hengsin: that's what I think a roadmap should have.
CarlosRuiz: and I think main goal is to support several clients
crooney: yes hengsin but that need not be in the roadmap
vpj-cd: the AJAX solve the issue with remote connection and we can be get indepndency of nomachine :-)
CarlosRuiz: not repeat the OpenBravo mistake dropping the swing client
crooney: yes the swing should stay
vpj-cd: yes openbravo are in top now , I think is because they are support to web client
CarlosRuiz: more Rod's words: "Look at individual issues and make small changes to resolve them – basically if it ain't broken don't fix it!"
hengsin: croo: why not ? I means to make sure everybody know the big picutre ...
vpj-cd: I think the new webclient is the based to TANG
crooney: hengsin: just that we can have it in the wiki but it's not the roadmap
hengsin: well, it is just a name, if you are not happy with it, call it something else, period.
CarlosRuiz: ok, returning to the 3-tier discussion
CarlosRuiz: my point is: I think ZK client can be done very fast - and we can have a really working web client
CarlosRuiz: and I think JSF+3-tier is very big, possibly for future versions
vpj-cd: yes I agree about jsp
vpj-cd: I think jsp do not solve our issue
hengsin: yes, next step is just web client, 3 tier is for future ...
teo_sarca: btw carlos, there is a lot of business logic in the "client" module... this should be moved in base or dbPort, and reuse it for ajax client
vpj-cd: we need AJAX to solve this issue
hengsin: and not to be confuse, jsf doesn't need 3 tier either.
vpj-cd: ok what option we have
crooney: what we want is the business logic in the middle tier so we can add many UI
vpj-cd: ZK, GWT, ECHO2, Seems
CarlosRuiz: or am I wrong? Maybe Vince didn't talk about 3-tier, maybe I read wrongly
vpj-cd: I think we only need a engine client
CarlosRuiz: what did you understand?
crooney: no he did ... he mentioned that TANG
vpj-cd: this should make work in the web browser
hengsin: yes, vince did but not to be confuse with his message, implementation of jsf client doesn't mandate a 3 tier architecture.
CarlosRuiz: ok
CarlosRuiz: so we can state - web client (initially non 3-tier)
CarlosRuiz: and in a later stage - 3-tier client (if really needed) "don't change anything for technology's sake"
vpj-cd: but this is easy if we think good the form to access the AD
hengsin: and we can have multiple experiement effort at this early/investigation state, as long as there are people who volunteer themselve. We can review later when those work are ready
CarlosRuiz: Kontro, we also have a thinking about TANG
CarlosRuiz: people can be confused too with TANG name -> next generation
CarlosRuiz: if TANG is a R&D lab - maybe the name must not state that it will be next generation - possibly will be, but is not sure
CarlosRuiz: but we see people in forums too saying - ok, we can wait until TANG is implemented, or when will be TANG?
hengsin: yes, tang is misleading, at least for now, as it may or may not be the 'next generation'
hengsin: we don't know yet.
crooney: or TANG R&D
CarlosRuiz: labs.adempiere.org :-)
CarlosRuiz: TANG Labs
vpj-cd: yes but the version 4 give some way
vpj-cd: this is very important to ppl or comunnity
CarlosRuiz: version 4?
hengsin: victor, there is no version 4, even 3.2 is not out yet :)
hengsin: we can
hengsin: we can't plan that far ahead.
vpj-cd: this is main reason when I set the 3.99 or 4 :-)
CarlosRuiz: this is moving so fast that we're having things not dreamt in roadmap
vpj-cd: but this was ide to begin
vpj-cd: yes but this dream give a way
hengsin: yes, victor, so why put a version number there for something that no one is sure of now.
vpj-cd: vs compiere
vpj-cd: how I told this is only a propouse
CarlosRuiz: yes, and is historical, I agree
vpj-cd: but generality in OS world the ppl need know the road map
CarlosRuiz: but the reality is moving other way
vpj-cd: yes if we need new road map is good
vpj-cd: but we need set and define
vpj-cd: what is way ?
vpj-cd: some know the way?
hengsin: yes, but most open source project only have roadmap for the next version, even for project like eclipse that have lots of resources ...
CarlosRuiz: exactly, that's the point, what are we dreaming today? and dreams must be open, not closed
CarlosRuiz: and setting up numbers and specific technologies close dreams
CarlosRuiz: and can stop dreamers :-)
vpj-cd: yes I agree
vpj-cd: then we need set the road map
vpj-cd: as a draft
vpj-cd: the community is very fast
hengsin: so we can have a section for items that is there for future investigation/enhancement ....
vpj-cd: is dynamic
CarlosRuiz: and we keep initial roadmap as an historical document - it's the spark
vpj-cd: we have not set posteria now we have here
vpj-cd: I think we can move the road map
vpj-cd: but this should give some way
CarlosRuiz: kontro, what do you think about TANG name?
vpj-cd: the ppl need know that adempiere are increase
vpj-cd: when we told about the your propuse to TANG
vpj-cd: is some idea
vpj-cd: the TANG have not in the roadmap
vpj-cd: this is the propouse the kontro
CarlosRuiz: yes, and I like the proposal, and the real hard work followed by kontro
CarlosRuiz: the issue is that name can be confusing for people, just like the closed roadmap
vpj-cd: the version 3.99 or $ is only an idea
vpj-cd: I think you can give to rich or modify
CarlosRuiz: Hengsin also told me about the numbering strategy as confusing
CarlosRuiz: people can't follow easily the odd/even numbering strategy
hengsin: haha ... we follow linux numbering scheme after they have drop it ...
vpj-cd: yes I agree
vpj-cd: but then now we need set the way
vpj-cd: and delete this confusion
CarlosRuiz: yes, we can think about a best strategy after 3.2
vpj-cd: ok the taslk about the 3.2 we have clear what is goal the 3.2
vpj-cd: ?
CarlosRuiz: kontro, AYT? we don't want to be rude with TANG
crooney: odd & even is very common? most users do not need to know the details but it is a quick way to see what kind of installation we have
vpj-cd: I think exist some type user
hengsin: most use just take bigger version number as the latest stable release :)
vpj-cd: the user that have a compiere in production
vpj-cd: and need solve issues very specific
hengsin: odd & even is for techie
vpj-cd: and other user that see to adempiere as a project very interest
vpj-cd: and they are interest by future in tegnology
crooney: prhaps not important now but in a few years it might simply things for us
crooney: if we do change the tang name it should be dual named for a while otherwise people will be saying .... what happended to tang? :)
vpj-cd: I think that tang is only one project
vpj-cd: ad posteria and libero
crooney: yes but do you want a user implementing a ERP?
vpj-cd: no is road map
vpj-cd: I think we need both
vpj-cd: because the user that are in ERP , they find new functionality no tecnical
vpj-cd: they need new feature in adempiere to inprove your process
vpj-cd: but also we get new developmnet
crooney: i mean if they can't figure out and odd & even numbering then there is not much hope for them implementing an enterprise wide app
vpj-cd: this new dev come if they can see a good road map with future
vpj-cd: to implementing a ERP you need a expertice in bisness process
vpj-cd: no only know of programing java or db :-)
vpj-cd: but I think adempiere give to all
CarlosRuiz: ok, but as I said this is a point that we can discuss after 3.2
vpj-cd: ok what do you think about 3.3?
crooney: yes and you also need technical skills
crooney: ok :)
CarlosRuiz: for today I think we can close the meeting, is too late in Malaysia, what do you think?
crooney: integrate all the big contributions ... I think that is a good plan
crooney: yes ok
hengsin: croo: what is the benefit of this odd and even scheme ? I can't think of any.
vpj-cd: Carlos I think we can set un draft to 3.3
vpj-cd: :-)
crooney: well when someone asks you in 5 or 6 years time to do a customisation for them and you ask what version and they say 3.1.5 then immediately you think hmmm that's not a stable version!
vpj-cd: but if you wnat we can talk this in the next meeting
CarlosRuiz: BTW, we're going to release 3.2 and immediately set trunk as 3.3?
vpj-cd: yes I think the trunk
hengsin: croo: how is that more informative than 3.1.0-b5 :)
vpj-cd: is development version but very stable
crooney: hengsin: ok touche .. I was thinking of 3.1.0 or 3.2.0 nobody mentioned extenions :)
kontro: Back here again. Had to fix neighbor's computer. Something wrong with reiserfs...
vpj-cd: yes Colin I can tell that 3.1.5 is very stable
crooney: but I guess from years of odd & even I have a distrust of odd numbered releases ... as I do for .0 releases :)
vpj-cd: I have from 1 march in production use postgresql with 60 user concurrent
CarlosRuiz: good news Victor, so we can release 3.1.6 as very stable release candidate
crooney: yes! with trunk or libero branch?
vpj-cd: I am use trunk
vpj-cd: the libero branch is the second stage
crooney: i might skip my planned XE phase then :)
vpj-cd: I have sing with mi customer the core
vpj-cd: adempiere
vpj-cd: but next stage is libero
crooney: great news then
CarlosRuiz: ok, Johannes is committing, I think we can set target for 3.1.6 this week, even tomorrow
crooney: i'll leave that to you in CC :)
CarlosRuiz: have someone time to help with the oracle seed?
vpj-cd: now postgresql is acting like a great mister
hengsin: by this week, that is.
vpj-cd: ok bye
trifon: what exactly with oracle sed?
vpj-cd: I find severt issue with 3.1.5 this wwek
hengsin: I'll do the postgresql seed.
vpj-cd: with shiment
CarlosRuiz: just prepare the oracle seed
vpj-cd: this issue corompe the onhand inventory
vpj-cd: I am fixed issue
CarlosRuiz: normally I import the previous release (3.1.5) - apply all the patches (including a new patch to set the version and date) - and export as reference for new seed 3.1.6
crooney: i was thinking at thye time of the last release .. is there a doc or howto for creating the releases? I have no idea what steps you take
vpj-cd: the problem is with validation credit
vpj-cd: Carlos do is complex if we release the change in AD also use 2pack?
crooney: with paying rather than crdeit sales victor? yes that is very annoying!
vpj-cd: the problem with credit validation is
vpj-cd: that stop shipment
vpj-cd: but the process substacting the onhand from mstorage
vpj-cd: but have not generate the transaction on mtransaction
vpj-cd: then the onhand is corrumped
crooney: oh! i didn't notice that... that's not good
CarlosRuiz: good catch Victor, is in trunk?
CarlosRuiz: Victor, didn't understand your question : release the change in AD also use 2pack?
vpj-cd: yes we change some things in 3.1.5
vpj-cd: the what happend if we generate a ad315.zip to this change use 2pack
vpj-cd: the any that are working with 3.1.4 only need applay the zip
vpj-cd: and you migration to 3.1.5 is ready
crooney: just looking at that problem with two users running the generate invoices (manual) at the same time and I got two invoices for the same sales order line
CarlosRuiz: so, the problem is solved but opened a new one (maybe the process must look for the status of sales order when generating the invoice - concurrency problem)
vpj-cd: but this is solve to pach of Low
crooney: oh hold on .. I was usnig the wromg window! ... old version! sorry
CarlosRuiz: but I think this new is low priority, at least 3 - it's difficult to happen
crooney: i think I need coffee :)
CarlosRuiz: :-)
crooney: yes agreed .. usually on one person in charge doing it
CarlosRuiz: I need lunch :-D
vpj-cd: Carlos you review my pach to replenish
vpj-cd: I can set into trunk?
vpj-cd: and my patch to MPeriod?
CarlosRuiz: which are those?
hengsin: victor, have u read my comment for your replenishment patch ?
CarlosRuiz: Victor, about your question on 2pack, I think the concept is ok, just not sure if 2pack manages well the entitytype and ID's for official dictionary entries
vpj-cd: yes Low but I undestand not what you need change
hengsin: http://sourceforge.net/tracker/index.php?func=detail&aid=1680304&group_id=176962&atid=879332
vpj-cd: what is you subgestion Low in the pach
vpj-cd: Carlos I import libro use 2pack and sequence the sytem
vpj-cd: I maked with you recomendation
vpj-cd: this working fine
CarlosRuiz: yes, but I mean, if you can export same sequences to other installations
hengsin: victor, you cannot just comment the delete, it is needed if replenishment type is not custom
CarlosRuiz: we must study better your recommendation Victor, I think that can be the way, just forget about system sequences
vpj-cd: but I think this way we can forget the comflict with sequences
CarlosRuiz: I think that can be the way, but we need to think and test more about that, it can open collateral problems
vpj-cd: ok Low this you line code
vpj-cd: wh.getReplenishmentClass() == null || wh.getReplenishmentClass().length ==
vpj-cd: 0 ) .
vpj-cd: where I need add or change
vpj-cd: I have the code in this moment
CarlosRuiz: and why not change the C_BPartner_ID to accept nulls, is a temporary table
CarlosRuiz: it will work?
vpj-cd: yes is easy solution
vpj-cd: yes I test with this solution
vpj-cd: but is more easy change the code that AD :-)
CarlosRuiz: not AD- is an alter table
vpj-cd: and I think that a ID should is 0 and not null
CarlosRuiz: but not in temp tables
crooney: all these strange bugs ... I'm sure glad we started Adempiere ... Inc would be years trying to find these ... we lots of "it works in the nighly build" message from kathy
CarlosRuiz: :-D but they don't want my "Directions" advice
trifon: i agree Colin.
vpj-cd: yes Carlos but this no is all the issue
CarlosRuiz: just again I was treated as a screamer throwing misinformation :-D
trifon: they even do not give me access to source code :(
vpj-cd: the issue is when you use the type replenish of type custom
CarlosRuiz: they are who are losing my advices ;-)
CarlosRuiz: hehehehehe
hengsin: I means if ( wh.getReplenishmentClass() == null || wh.getReplenishmentClass().length == 0 ) { // the original delete code goes here }
CarlosRuiz: why the delete is bad?
vpj-cd: ok low you say wh.getReplenishmentClass() == null
CarlosRuiz: it must delete zero records normally
vpj-cd: sorry Delete rows where nothing to order
vpj-cd: but I need this logic to custom type
vpj-cd: I need delete the qtytoOrder also to custom type
vpj-cd: my current problem is with current code
vpj-cd: this delete all record before the caculate my algoritm
vpj-cd: then I get empty report
vpj-cd: is this reason I move this code in last section
vpj-cd: this way working fine to any type replenish
CarlosRuiz: you mean after you calculate your algorithm
vpj-cd: yes
CarlosRuiz: let me understand
CarlosRuiz: you have a custom process that fills the T_Selection previously, then you call this process and is deleting your previous work
vpj-cd: yes but this issue
vpj-cd: delete all the record
vpj-cd: before I can calculate the qtytoorder
vpj-cd: this is error
CarlosRuiz: you customized ReplenishReport directly or is another external process?
vpj-cd: I use custom class
vpj-cd: and implement the interface repleanish
vpj-cd: I set my class in the warehouse windows
vpj-cd: the process call when of type is custom
CarlosRuiz: oh I see, you have other rules to calculate QtyToOrder
vpj-cd: yes I use the rule of my customer
vpj-cd: my rule set qtyto orden
vpj-cd: when onhand is < that qtyminimal
CarlosRuiz: can you put that rule before the delete?
vpj-cd: then I have get new qtyorder based in qty maximal
vpj-cd: current the custom class are in finish
vpj-cd: the methos filltable
CarlosRuiz: I think the delete is needed but after (at the end) of the fillTable - is that what you're saying?
CarlosRuiz: is needed to not generate orders with qty = 0
vpj-cd: yes
vpj-cd: I move this code and set in the last section
CarlosRuiz: ok, looks ok, what's the hengsin suggestion?
hengsin: insted of commenting the delete code, add a if ( wh.getReplenishmentClass() == null || wh.getReplenishmentClass().length == 0 ) check
CarlosRuiz: on the other hand, I think updated/updatedby are not needed in temp tables
CarlosRuiz: temp tables are not intended to have model classes
vpj-cd: yes Carlos but the PO in postgresql faild
vpj-cd: then the transaction is rollback
CarlosRuiz: which PO
vpj-cd: in Low but if I only wt your code
vpj-cd: then I need set another code to delete the qtytoorder in zero
vpj-cd: yes with PO this lauch a error and the transaction is faild and rollback
vpj-cd: then I have not update record
CarlosRuiz: Victor the issue here is that is very difficult for us to review a problem without a reproducible test case
CarlosRuiz: we can't know what is trying to solve, and if is solving it really, and if it's opening other collateral problems
vpj-cd: on I test with custom and normal type replenish
vpj-cd: and all working fine
vpj-cd: but if you want wait another can test
vpj-cd: is good to me
vpj-cd: I have soleve my issue :-0
vpj-cd: ;-)
CarlosRuiz: yes, and I trust your testings, the problem is that is a customization problem that can't be reproduced in GardenWorld (at least easily without the customized code)
CarlosRuiz: what do you think on this hengsin?
vpj-cd: and Low test and are agree
vpj-cd: with you commnet
CarlosRuiz: yes, I'm not against the change, just that I can't confirm if it works easily, looking around the code looks ok for me
vpj-cd: then somebody need test with type replanish
CarlosRuiz: but you tested
hengsin: no, I didn't test, only comment base on your patch. victor, can u upload your final patch for review ?
crooney ha salido de la sala (quit: Read error: 110 (Connection timed out)).
crooney [n=crooney@89.19.76.206] ha entrado en la sala.
CarlosRuiz: ok, gtg
vpj-cd: Low the final pach is the class I attached in the sf
crooney ha salido de la sala (quit: Client Quit).
CarlosRuiz: Víctor, also where is the PO broken (updated/updatedby) ?? is in customized code or is reproducible in GardenWorld with postgres?
vpj-cd: yes
hengsin: for that patch, I don't think the comment of the delete is good.
vpj-cd: the issue is when you use custom type
vpj-cd: because you use the X_T_Replenish
vpj-cd: when you use the normal type
vpj-cd: have not use the model
vpj-cd: Low I do not deleted the code
vpj-cd: only move the last section
CarlosRuiz: and why not use the X_T_Replenish in customized code?
CarlosRuiz: I mean avoid usage of X_T_Replenish in customized code
vpj-cd: no
vpj-cd: this code is the normal
vpj-cd: code X_T_Replenish replenish = replenishs[i];
vpj-cd: if (replenish.getReplenishType().equals(X_T_Replenish.REPLENISHTYPE_Custom))
vpj-cd: {
vpj-cd: BigDecimal qto = null;
vpj-cd: you can see in the process
vpj-cd: this is use from compiere
vpj-cd: when you choose type custom
vpj-cd: line 348
hengsin: victor, // Custom Replenishment
hengsin: String className = wh.getReplenishmentClass();
hengsin: if (className == null || className.length() == 0)
hengsin: return;
hengsin: isn't the delete code wouldn't be executed for standard replenishment ?
CarlosRuiz: you're right hengsin
hengsin: so you need it at both place.
vpj-cd: yes when you use you our class the replenish have fiald
vpj-cd: faild
CarlosRuiz: and Victor, you're right too, it's a big bug, X_T_Replenish will fail, I suppose in oracle too
CarlosRuiz: replenish.save
CarlosRuiz: or drop the return and put brackets around customreplenish code
CarlosRuiz: and negate the if obviously
hengsin: yes, that should work too
vpj-cd: ok my another
vpj-cd: http://sourceforge.net/tracker/index.php?func=detail&aid=1680306&group_id=176962&atid=879332
vpj-cd: what do you think
CarlosRuiz: yes, I was reviewing and looks of for me, strange thing with Mac
vpj-cd: yes the issue is with time stamp
vpj-cd: I review in postgresql forum
vpj-cd: and they say that exist some issue with some JVM
vpj-cd: this case is mac
vpj-cd: I use same logic that other process
hengsin: sounds like jvm bug
vpj-cd: I test this issue with windows client and linux and mac linet
hengsin: Apple JVM is very buggy ...
vpj-cd: client and now are working fine
CarlosRuiz: does Adempiere run with non-sun JVM's ??
hengsin: not tested, but we can support those without maintenaner
hengsin: we can't
vpj-cd: in windows I use jvm from sun
vpj-cd: also in linux
vpj-cd: and mac I use apple
hengsin: maybe some problem in TimeUtil ?
vpj-cd: Low the issue is the syntaxis of timestamp with time zone
vpj-cd: Apple return of type that when you set as parameter in db postgresql ignore when you use BETWEEN
vpj-cd: the issue is you have not get a resultset
vpj-cd: then the logic fail
vpj-cd: I change this logic to TO_DATE
vpj-cd: and convert change to TO_TIMESTAMP in postgresql this trunk of the date and now working fine with any client
CarlosRuiz: ok, gtg
CarlosRuiz: have a meeting with a customer
CarlosRuiz: thank you very much
hengsin: bye all
CarlosRuiz: bye
hengsin ha salido de la sala.
